Lyft
Machine Learning•2026-02-12
Trusting the Untestable: Validation and Diagnostics for the Doubly Robust Models
This post from Lyft explains how they validate and diagnose Doubly Robust (AIPW) models used for causal inference when A/B testing is not feasible.
